The Epstein-Barr virus (EBV) is a member of the herpes virus family and is the causitive agent of infectious mononucleosis (IM) in humans. EBV has also been implicated in the pathogenesis of Burkitt's lymphoma, nasopharyngeal carcinoma, and B lymphocyte neoplasms arising in immunosuppressed patients. Circumstantial evidence has also indicated a possible role for this virus in human autoimmune disease such as rheumatoid arthritis and Sjogren's Syndrome.
EBV is an extremely common environmental agent infecting 80-100 percent of the individuals around the world. The initial or primary infection may be acute or sub-clinical. This is followed by a long period during which the EBV infection is latent in B lymphocytes present in the circulating blood, lymph nodes, and spleen.
Latency is the process by which a virus is present intracellularly in an unexpressed or partially expressed state. This latency can be reactivated. Although the host factors that control latency in vivo are poorly known, there is some evidence to suggest that failure of one or more immune mechanisms is an important factor.
Cytotoxic and suppressive T cell elements of the immune response to EBV are reported to be very important in suppressing acute infection by EBV in IM. They are also important in prohibiting the uncontrolled outgrowth of B lymphocytes latently infected with EBV.
Failure of T cell suppressor mechanisms is thought to be important in allowing the emergence of African Burkitt lymphoma, nasopharyngeal carcinoma, B cell lymphomas arising as a consequence of immunosuppressive therapy used to prevent rejection of organ transplantation, and lymphomas as arising during treatments of various auto immune disorders. Epstein and Achong eds., " The Epstein-Barr Virus. ", Spring-Verleg, Berlin, Heidelberg (1970); and Crawford et al., Lancet, 1355 (1980). In addition, the failure of these T cell mechanisms and consequent overgrowth of EBV-infected lymphocytes is thought to play a role in reheumatoid arthritis. Slaughter et al., J. Exp. Med., 148:1429 (1978); Depper et al., J. Immunology, 127:1899 (1981) and Tosato et al., N. Engl. J. Med. 305:1238 (1981).
The serological and cell-mediated immune responses that follow primary infection by EBV are well documented and reflect the host's response to the viral antigens expressed during the course of infection. The profile of these responses as well as the detection of the antigens in tissues are becoming increasingly useful in the diagnosis of EBV-associated diseases.
The earliest EBV-associated antigen that can be detected after infection is EBV-induced nuclear antigen (EBNA). EBNA has been detected in the nucleus of latently-infected growth-transformed B lymphocytes. EBNA has also been detected in the nuclei of African Burkitt tumor lymphoblasts and anaplastic nasopharyngeal carcinoma cells.
The concentration of EBNA in cell nuclei of EBV-infected By lymphocytes fluctuates during various phases of the cell's reproductive cycle. Thus, it is believed EBNA is cyclically being synthesized and degraded. As a result of such degradation, protein fragments (polypeptides) of EBNA traverse the cellular cytoplasm and are believed to exist or be expressed on the outer membrane. However, specific EBNA degradation polypeptides have not been identified to date.
It is believed that while in or on the outer cell membrane, EBNA degradation polypeptides constitute a significant stimulus to the host's T lymphocytes and initiate the immune response that results in the production of anti-EBNA antibodies. It is also believed that the specific T cell response to B cells expressing EBNA degradation polypeptides on their surfaces may contribute to the generation of cytotoxic and suppressive T cells important in restricting growth of EBNA-containing (EBV-infected) B lymphocytes.
Thus, assays for the presence of both EBNA and anti-EBNA antibodies are of importance in several common clinical situations. In addition, a vaccine against EBV-infected B lymphocytes would also be of clinical importance.
Anti-EBNA antibodies are typically assayed using the tedious anti-complement immunofluorescence technique (ACIF). Reedman et al., Int. J. Cancer, 11:499-520 (1973). This assay involves fixing EBV-transformed human B cells to a microscope slide. Various dilutions of a patient's serum are then added to the fixed cells. Because anticomplementary sera may yield false-negative reactions or prozones when they are mixed with the complement (a two-stage procedure), it is essential to charge the test cell smears consecutively with serum, complement, and the anticomplement-fluorescence conjugate (a three stage procedure).
There are several problems with this assay. These include the fact that the assay is relatively insensitive and requires amplification mediated through complement. In addition, this assay is not entirely specific and may not be interpreted in patients whose serum contains antibodies to mammalian cell nuclei. Still further, quantitative results obtained using an anti-complement immunofluorescence assay are difficult to reproduce. As a consequence of these and other reasons, assays for anti-EBNA antibodies have generally been confined to a few, specialized laboratories.
The above difficulties in assaying for anti-EBNA antibodies stem from the lack of relatively pure EBNA. Purification of EBNA from mammalian cell tissue cultures is complex because of the antigen's low concentration and polymorphology. Although it is easier and less costly to use whole cells expressing EBNA, as in the current technique, the problems of specificity and reproducibility are directly tne result of using whole cells.
Genetic engineering and synthetic polypeptide technologies have recently provided solutions to the problem of manufacturing large quantities of protein and polypeptide antigens. However, both techniques are effective only if the amino acid residue sequence of the native protein is known.
The amino acid residue sequence of a natural protein can be determined from the protein itself, but this is often difficult. The gene nucleotide sequence that codes for the protein may also reveal the protein's amino acid residue sequence. However, all DNA sequences have three possible reading frames, each of whicn yields a different protein. Therefore, the correct reading frame must be known to deduce the correct amino acid residue sequence of a protein from its gene.
The correct reading frame of a DNA sequence coding for a protein, and therefore the protein's amino acid residue sequence, may be determined through the use of antibodies. This strategy involves manufacturing an array of protein fragments or polypeptides whose amino acid residue sequences correspond to the sequences obtained from the three possible gene products. The protein fragments or polypeptides that induce antibodies that immunoreact with the gene's natural protein product thereby indentify the gene's correct reading frame. Conversely, if antibodies to the natural protein recognize the manufactured protein fragments or polypeptides, the relationship between gene and protein is also established.
Heller, et al., J. Virol., 44:311-320 (1982), reported the DNA sequence for a portion of the EBV genome that was found to contain an internal region, designated IR3, consisting of direct repeats of a hexanucleotide and two nonanucleotide sequences. They cited evidence suggesting that the sequence surrounding and including IR3 contains the gene coding for EBNA. However, since it was not known which of the three possible DNA sequence reading frames was translated, Heller, et al., supra, were not able definitely to deduce the amino acid sequence for the possible EBNA protein.
In September 1983, Hennessy and Kieff, Proc. Natl. Acad. Sci., U.S.A., 80:5665-5669 (1983), reported establishing the natural reading frame for the EBV DNA sequence reported by Heller, et al., supra. Essentially, they isolated IR3 DNA, cleaved it into small random pieces and inserted the pieces into the lacZ gene of an E. coli expression vector such that all three EBNA gene reading frames were expressed, each in a different clone. The lacZ gene codes for beta-galactosidase, a bacterial enzyme. The IR3-lacZ gene fusion product is expressed in E. Coli as a fusion protein with the IR3 protein sequence inserted between amino acids 7 and 9 (8 being deleted in the construction process) of the beta-galactosidase protein molecule.
Hennessy and Kieff, identified an IR3-lacZ gene fusion that was expressing IR3 DNA in its natural reading frame by screening for fusion proteins that were recognized by anti-EBNA positive human sera. A plasmid so identified was designated pKH182-44.
To confirm that the protein expressed by pKH182-44 contained EBNA-specific antigenic determinants, Hennessy and Kieff, supra, raised antisera in rabbits against cyanogen bromide-cleaved (CNBr) IR3-galactosidase fusion protein. The CNBr fragment used as an immunogen contained 53 amino acids homologous to EBNA and 89 amino acids homologous to beta-galactosidase. These antisera recognized natural EBNA in EBV-infected cells using indirect immunofluorescence.
The results of Hennessy and Kieff appear to be dependent on the repetitive nature of the EBNA IR3 domain. The fusion protein produced by pkH182-44 contains a relatively long segment homologous with the IR3 domain (e.g. 53 amino acids). It is, therefore, not surprising that the fusion protein and CNBr fragment thereof contained antigenic determinants. Furthermore, Hennessy and Kieff did not identify which of the sequences repeated in their fragment were acting as antigenic determinants.
Although Hennessy and Keiff were able to genetically manufacture a material recognized by anti-EBNA antibodies in human serum, it would be cumbersome to use in a clinical setting because of its design. The 53 amino acid residue segment of their fusion protein that is homologous to EBNA is physically and chemically part of the beta-galactosidase protein. Its immununological properties are, therefore, influenced by those portions of the beta-galactosidase molecule from which it cannot be separated. In fact, all of the human sera used in their study reacted with beta-galactosidase, and required treatment with beta-galactosidase to adsorb and remove this reactivity before testing for specificity against the genetically manufactured protein.
Another approach to the interrelated problems of determining a gene's correct reading frame and manufacturing large quantities of pathogen-related antigens (immunogens) for clinical and diagnostic purposes is the use of synthetic polypeptide chemistry. This method of manufacturing antigens (immunogens) has an advantage over the genetic engineering methods described above. Synthetic polypeptide antigens do not contain natural protein by-products or fragments thereof, and thereby their use eliminates the possibility of unwanted cross reactivity and the need to pretreat serum samples as in the Hennessy and Kieff study.
While the general concept of preparing synthetic antigens (immunogens) and using them to induce antibodies of predetermined specificity has been described, there remains a large area of this technology that continues to defy predictability. There are at least two reasons for this. First, a synthetic antigen (immunogen) does not necessarily induce antibodies that immunoreact with the intact protein in its native environment. Second, a host's natural antibodies to a naturally occurring immunogen such as a viral protein rarely immunoreact with a polypeptide that corresponds to a short linear portion of the immunogen's amino acid residue sequence. This latter phenomenon is believed to be the result of short linear polypeptides lacking required secondary and tertiary conformational structures.
Much of the work on the binding of peptide by antibody made to proteins is summarized in a review by Benjamini, E., et al., Current Topics in Microbiology and Immunology 58:85-134 (1972). The role of peptide structure in antibody binding has been emphasized by Goodman, J. W., Immunochem 6:139-149 (1969).
Most of the studies concerned with how changes in the sequence of peptides effect antibody binding have been interpreted as indicating tnat tne structure of tne antibody combining site plays a predominant role. The effect of sequence and structural changes in these studies is intermixed and difficult to segregate. Some of these studies can equally well be explained by structural changes in antigen effecting the binding.
Antibody response at the molecular level involves binding of an antigen of defined sequence (primary structure) and in a defined conformation (secondary and tertiary structure). Immune response to protein antigens has traditionally been interpreted as being directed against primary, secondary or tertiary structue of the protein.
This classification scheme may have some validity for proteins tnat have a well defined overall structure at physiological temperatures and solutions. However, its validity is in doubt for peptide antigens that nave a more dynamic structure.
Several groups have reported structural studies of polymers of repeating sequence of glycine and alanine or glycine and serine that were synthesized as models of silk fibroins [Anderson et al., J. Mol. Biol. 67:459-468 (1972)] and collagen [Anderson et al., BBRC 39:802-808 (1970); Doyle et al., J. Mol. Biol. 51:47-59 (1970]. The most systematic study has been that of Brack et al., Biopolymers 11:563-586 (1972) who reported synthesis of a series of block homopolymeric polypeptides in which the homopolymeric block repeating units had the formula (Ala.sub.x -Gly.sub.y) wherein when x=1, y=1, 2 and 3; when x=2, y=1, 2 and 3; and when x=3, y=3.
The results reported from this latter study were that in the solid state the block homopolymers composed mostly of alanine were alpha-helical, and those containing mostly glycine were disordered. In solution, polyalanine was reported to be alpha-helical, but poly-(Ala.sub.2 -Gly.sub.1) was reported to be in beta-antiparallel form. The more glycine-rich polymers were said to have another fixed structure that was reported as neither alpha-helix nor the beta-structure.
The homopolymerized blocks of glycine and alanine reported by Brack et al. were prepared by the condensation polymerization of the di- through hexapeptide repeating units having a carboxyl-terminal glycyl residue in active ester form. Degrees of polymerization from 2 through 68 were reported for poly(Ala-Gly.sub.2).
Even though the solvents used in those studies were not physiologically acceptable, e.g., water or phosphate buffered saline, the results illustrate two points: (1) structural changes can occur as the sequence of a polypeptide changes, and (2) structural changes also occur during the transition from solution to solid state.